Studies on microcirculation in insulin resistance
Josefin Olausson
Department of Molecular and Clinical Medicine Institute of Medicine
Sahlgrenska Academy at University of Gothenburg
Gothenburg 2015
Cover illustration by Madelen Lindgren, Madebylen
Studies on microcirculation in insulin resistance
© Josefin Olausson 2015 josefin.olausson@wlab.gu.se ISBN 978-91-628-9609-6 (Print) Printed in Gothenburg, Sweden 2015 Ineko AB
Mormor och Morfar
resistance
Josefin Olausson
Department of Molecular and Clinical Medicine, Institute of Medicine Sahlgrenska Academy at University of Gothenburg
Göteborg, Sweden ABSTRACT
The overall aim of this thesis was to investigate the microcirculation in insulin resistance, with focus on the expression of endothelin-1, through a translational approach.
Specific aims: 1) Investigate if circulating endothelin-1 levels predicts incident coronary heart disease events. 2) To assess if sex differences modify endothelin-1 as a predictor of type 2 diabetes. 3) To investigate if microvascular insulin resistance impairs insulin delivery to the subcutaneous adipose tissue and skeletal muscle. 4) To investigate if acute administration of the PDE-5 inhibitor tadalafil induces positive vascular, metabolic and anti-inflammatory effects in type 2 diabetes. 5) To further elucidate the molecular action of tadalafil in tumour necrosis factor-α (TNF-α) stimulated human endothelial cells.
Principal findings: The population-based cohort in Vara-Skövde was investigated for paper I- II. During baseline cardiovascular risk factors and endothelin-1 were assessed and incident coronary heart disease (CHD) was followed-up during a 10-year period (paper I). Endothelin- 1 levels had a predictive value for incident CHD in women, but not in men. A randomly selected subgroup was investigated in a follow-up after 10 years, and impaired glucose tolerance (IGT) and T2D was documented for paper II. Here, higher quartiles of endothelin-1 at baseline were associated with IGT/T2D at follow-up in women. Paper III investigates microvascular aspects of insulin resistance using microdialyis; participants with T2D and age- matched healthy controls were studied after an oral glucose load. Participants with T2D had decreased delivery of insulin to adipose tissue, and a blunted subcutaneous adipose tissue blood flow compared with controls. In paper IV, T2D participants received either placebo or tadalafil (20 mg) before a mixed meal in a randomized controlled trial. Tadalafil increased forearm blood flow, glucose uptake and capillary recruitment, and blunted a postprandial increase of endothelin-1. In paper V, the effects of tadalafil were studied in an experimental setting using TNF-α stimulated endothelial cells. Tadalafil treatment decreased expression of c-Jun N-terminal kinase (JNK) phosphorylation as well as reduced gene expression and secretion of endothelin-1.
Conclusions: This thesis shows that (i) endothelial dysfunction precedes IGT/T2D and CHD, and that endothelin-1 may pose as a risk factor for women, (ii) delivery of insulin from the circulation to subcutaneous adipose tissue is impaired in participants with T2D, and that participants with T2D exhibit a blunted postprandial blood flow response, (iii) acute administration of tadalafil induces positive vascular and metabolic effects in the postprandial state in T2D, and tadalafil decrease gene expression of endothelin-1 in cultured endothelial cells by decreasing activation of JNK.
Keywords: endothelin-1, coronary heart disease, type 2 diabetes, phosphodiesterase-5 inhibition, tadalafil,
Antalet personer med typ 2 diabetes (åldersdiabetes) ökar i världen. Denna ökning tros bero på dagens livsstil med mycket stillasittande och att fler personer idag är överviktiga.
Hjärtsjukdom är en av de vanligaste dödsorsakerna till följd av diabetes. Typ 2 diabetes definieras som en okänslighet mot insulin som är ett hormon som bildas i bukspottskörteln och som gör att kroppen kan ta upp socker från blodet. När man får en okänslighet mot insulin, bildar kroppen större mängder av insulin efter en måltid för att upprätthålla sockerupptaget.
Insulinet är även viktigt för att styra hur kärlen beter sig, genom att stimulera de celler som sitter på insidan av kärlen, endotelcellerna, att bilda olika ämnen. Två av dessa ämnen är speciellt viktiga, kväveoxid (NO) och endothelin-1. Dessa ämnen är principiellt varandras motsatser, NO gör så att kärlen slappnar av medan endothelin-1 gör så att kärlen drar sig samman. Vid typ 2 diabetes är signalen som gör att NO bildas av insulin skadad, medan signalen som bildar endothelin-1 fortfarande fungerar. Detta gör att framför allt de små kärlen blir mer sammandragna och man får inte lika bra blodflöde i muskel- och fettvävnad, vilket medför att sockret inte når sin målvävnad.
I denna avhandling har vi tittat på nivåer av endothelin-1 i blodet hos 2816 personer som bor i Vara och Skövde, dessa har sedan följts upp för att se vilka som fick hjärtsjukdom under en 10-års period. Ungefär en tredjedel kom tillbaka på ett nytt besök där de undersöktes utefter insulinkänslighet och ifall de fått typ 2 diabetes. Denna studie visade att de kvinnor som hade höga nivåer av endothelin-1 vid grundundersökningen hade större risk för hjärtsjukdom, samt en större risk för att utveckla en okänslighet mot insulin och typ 2 diabetes. Det syntes dock ingen ökad risk hos männen, vilket fick oss att dra slutsatsen att det finns en skillnad mellan könen gällande effekterna av endothelin-1 och utvecklingen av dessa sjukdomar.
I en annan studie jämfördes personer med typ 2 diabetes och friska kontroller genom att mäta insulin-nivåerna i underhudsfettet och i en underarmsmuskel. Där såg vi att insulinet inte kom fram till underhudsfettet från blodet hos personer med typ 2 diabetes lika snabbt som hos friska, och att deras kärlåterhämtning var sämre. Detta ger stöd åt teorin att kärlens endotelceller aktivt transporterar insulinet till sin målvävnad.
Tadalafil är ett läkemedel som efterliknar effekterna av NO genom att öka ämnet cGMP inuti cellerna. Tadalafil har tidigare visats hjälpa endotelcellernas funktion hos personer med impotens, och därför ville vi undersöka ifall denna positiva effekt även sker hos personer med typ 2 diabetes. För att göra detta gav vi forskningspersonerna antingen tadalafil eller placebo och mätte sockerupptag och blodflöde i en muskel i underarmen. Vi mätte även endothelin-1 och inflammationsmarkörer i blodet hos forskningspersonerna. Vi fann att de personer som fick tadalafil hade ett förbättrat sockerupptag och ökat blodflöde i de små kärlen i muskeln efter måltiden, samt att den ökning av endothelin-1 efter måltid som sågs hos den grupp som fick placebo inte sågs i gruppen som fick tadalafil.
För att ytterligare fördjupa vår kunskap om de mekanismer som tadalafil påverkar inuti cellen odlades endotelceller på laboratoriet. Genom att behandla cellerna med tadalafil samtidigt som man stimulerade dem med en inflammatorisk markör kunde vi mäta deras uttryck av gener och proteiner. Även här minskades produktion och utsöndring av endothelin-1, dessutom minskades genuttryck av inflammationsmarkörer. Vi kunde även identifiera ett protein som minskas av tadalafil, c-Jun N-terminal kinase.
Sammanfattningsvis förevisar detta vikten av ett fungerande endotel vid utvecklingen av diabetes och hjärtsjukdom, framför allt hos kvinnor. Dessutom kan läkemedlet tadalafil
LIST OF PAPERS
This thesis is based on the following studies, referred to in the text by their Roman numerals. * Contributed equally.
I. Bledar Daka*, Josefin Olausson*, Charlotte A Larsson, Margareta I Hellgren, Lennart Råstam, Per-Anders Jansson, Ulf Lindblad.
Circulating concentrations of endothelin-1 predict coronary heart disease in women but not in men – a longitudinal observational study in the Vara-Skövde Cohort. Accepted for publication in BMC Cardiovascular Disorders 2015.
II. Josefin Olausson*, Bledar Daka*, Margareta I Hellgren, Charlotte A Larsson, Max Petzold, Ulf Lindblad, Per-Anders Jansson.
Endothelin-1 as a predictor of impaired glucose tolerance and type 2 diabetes – A longitudinal study in the Vara-Skövde Cohort.
Submitted.
III. Josefin Olausson, Reza Mobini, Per Fogelstrand, Karin Mossberg, Emanuel Fryk, Lena Strindberg, Lillemor Mattsson Hultén, Per- Anders Jansson. Delivery of insulin to subcutaneous adipose tissue and skeletal muscle in type 2 diabetes patients and healthy controls – A microdialysis study. Manuscript
IV. Lovisa Sjögren, Josefin Olausson, Lena Strindberg, Reza Mobini, Per Fogelstrand, Lillemor Mattsson Hultén, Per-Anders Jansson.
Postprandial effects of the phosphodiesterase-5 inhibitor tadalafil in people with well-controlled type 2 diabetes mellitus: A randomised controlled trial. Accepted for publication in Diabetic Medicine 2015.
V. Josefin Olausson*, Lovisa Sjögren*, Reza Mobini, Emanuel Fryk, Per Fogelstrand, Lillemor Mattsson Hultén, Per-Anders Jansson.
Tadalafil decreases expression of endothelin-1 in TNF-α-activated human endothelial cells – possible role of the c-Jun N-terminal Kinase pathway. Manuscript.
ABBREVIATIONS ... IV
1 INTRODUCTION ... 1
1.1 Type 2 diabetes ... 1
1.2 Coronary heart disease ... 3
1.3 Microcirculation ... 3
1.4 Insulin resistance and endothelial dysfunction ... 4
1.4.1 Endothelin-1 ... 6
1.4.2 Obesity and inflammation ... 8
1.5 Insulin signalling and action ... 8
1.5.1 Transendothelial transport ... 10
1.5.2 Postprandial vascular response ... 11
1.6 Phosphodiesterase-5 inhibition ... 12
1.7 The gender aspect ... 13
1.8 Endothelin-1 as a biomarker and type 2 diabetes heredity ... 14
2 AIMS ... 16
3 PARTICIPANTS AND METHODS ... 17
3.1 Paper I and II ... 17
3.1.1 Participants ... 17
3.1.2 Study procedure ... 17
3.2 Paper III and IV ... 19
3.2.1 Participants ... 19
3.2.2 Study procedure ... 19
3.2.3 Microdialysis technique ... 21
3.2.4 Blood flow measurements ... 22
3.2.5 Assessment of endothelial function ... 23
3.3 Paper V ... 25
3.3.1 Cell culture and Experimental Design ... 25
3.4 Statistical analysis ... 26
4 SUMMARY OF RESULTS ... 29
4.1 Paper I – Circulating endothelin-1 levels predict coronary heart disease ... 29
4.1.1 Study design ... 29
4.1.2 Main results ... 29
4.2 Paper II – Endothelin-1 as a predictor of type 2 diabetes ... 30
4.2.1 Study design ... 30
4.2.2 Main results ... 30
4.3 Paper III – Delayed delivery of insulin to adipose tissue in type 2 diabetes ... 31
4.3.1 Study design ... 31
4.3.2 Main results ... 31
4.4 Paper IV – Postprandial effects of the PDE-5 inhibitor tadalafil ... 31
4.4.1 Study design ... 31
4.4.2 Main results ... 32
4.5 Paper V – Effects of tadalafil in TNF-α stimulated endothelial cells ... 32
4.5.1 Study design ... 32
4.5.2 Main results ... 32
5 DISCUSSION ... 35
5.1 Major findings ... 35
5.1.1 Endothelin-1 as a predictor of coronary heart disease and type 2 diabetes ... 35
5.1.2 Delivery of insulin to adipose tissue and skeletal muscle ... 37
5.2 Role of tadalafil in vivo and in vitro ... 39
5.3 Strengths and limitations of the study ... 42
6 SUMMARY AND CONCLUSION ... 43
7 FUTURE PERSPECTIVES ... 44
8 ACKNOWLEDGEMENTS ... 46
9 REFERENCES ... 48
ACE Angiotensin converting enzyme AP-1 Activator protein-1
Apo Apolipoprotein
ATBF Adipose tissue blood flow BMI Body mass index
cGMP Cyclic guanosine monophosphate CHD Coronary heart disease
c-PTIO 2-(4-Carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3oxide potassium salt
CRP C-reactive protein DPP Dipeptidyl peptidase EBM Endothelial basal medium ECE Endothelin-converting enzyme eNOS Endothelial nitric oxide synthase ELISA Enzyme-linked immunosorbent assay ERK Extracellular regulated kinase ETA Endothelin receptor A ETB Endothelin receptor B FFA Free fatty acids
FSH Follicle stimulating hormone GLP Glucagon-like peptide GLUT4 Glucose transporter 4 HbA1c Glycated haemoglobin HDL High-density lipoprotein
HOMA-IR Homeostatic model assessment of insulin resistance HUVEC Human umbilical vein endothelial cell
IAUC Incremental area under the curve ICAM-1 Intercellular adhesion molecule-1 I/C Interstitial/Circulating
ICD International statistical classification of diseases and related health problems IFG Impaired fasting glucose
IRS Insulin receptor substrate IκBα Inhibitory kappa B alpha IKK IκB kinase
JNK c-Jun N-terminal kinase LDL Low-density lipoprotein MAPK Mitogen activated protein kinase NF-1 Nuclear factor-1
NFκB Nuclear factor kappa B NO Nitric oxide
ODQ 1H-[1,2,4]Oxadiazolo[4,3-a]quinxolain-1-one OGTT Oral glucose tolerance test
ONOO Peroxynitrite
PI3K Phosphoinositide 3 kinase PDE-5 Phosphodiesterase-5 PKG Protein kinase G
RHI Reactive hyperemia index ROS Reactive oxidative species sGC Soluble guanylate cyclase Shc SH2-domain containing protein siRNA Small interfering RNA T2D Type 2 diabetes mellitus TNF-α Tumour necrosis factor alpha VCAM-1 Vascular cellular adhesion molecule-1 WHR Waist-hip ratio
133Xe 133Xenon
1 INTRODUCTION
Insulin resistance is a common denominator for numerous pathologies. It is foremost described in the area of cardiovascular diseases, however mechanisms causing insulin resistance are complicated and several are still not completely elucidated. The microcirculation is heavily implicated in the progress of insulin resistance, and a close relationship has been described between microvascular dysfunction and insulin resistance. This thesis focuses on the potential predictive value of a marker for microvascular dysfunction, endothelin-1, investigates the transendothelial transport of insulin, and explores the effect of a drug, tadalafil, which might have positive effects on the microcirculation.
1.1 Type 2 diabetes
Type 2 diabetes mellitus (T2D) is considered one of the major health concerns in the western world, with an increasing prevalence [1]. Diabetes is the fifth most common cause of death globally [2], and cardiovascular disease is the principal cause of death among the complications of T2D [3]. The growing prevalence of diabetes, with an estimation of 285 million persons affected worldwide, has been attributed to a more sedentary lifestyle in combination with an excessive dietary intake leading to an increasingly obese population [4]. Additional risk factors for T2D include smoking, hypertension and a genetic predisposition for diabetes [4, 5]. T2D is characterised by a peripheral insulin resistance defined by an impaired insulin-mediated whole body glucose-uptake and, in later stages, impaired insulin secretion [6, 7]. The pathophysiology of T2D involves several tissues in the body, in the pancreas the glucagon production by α-cells is increased due to impaired glucose sensing, and β-cell loss of function leads to decreased insulin production [8, 9]. The brain is involved through a compromised appetite regulation and insulin resistance through impaired insulin signalling in hypothalamus [10], the liver by increased glucose production and triglyceride secretion and the kidney by an increased glucose reabsorption [8, 9]. Further, skeletal muscle and the adipose tissue displays decreased glucose uptake, and the adipose tissue has an increased lipolysis. The gastrointestinal tract has an impaired incretin function [11], and the intestinal microbiome has during recent years emerged as an important factor in the pathophysiology of T2D [12]. The cardiovascular system is also affected in T2D, with a compromised micro- and
Figure 1. A schematic overview of various risk factors and pathophysiology involved in type 2 diabetes
1.2 Coronary heart disease
Coronary heart disease (CHD) most frequently causes death among the cardiovascular diseases [13]. CHD refers to different conditions of failing circulation of the heart and includes coronary atherosclerosis, myocardial infarction and angina pectoris [14], among others. The risk of CHD is doubled in subjects with T2D [15], and other conditions implicated in CHD are hypertension [16] and obesity [17]. Endothelial dysfunction has been suggested to play a central role in atherosclerosis [18, 19]. Endothelial dysfunction has also been shown to increase the risk for cardiovascular events synergistically with T2D, impaired glucose metabolism or insulin resistance [20]. However, CHD is a multifactorial condition and several other factors are implicated in the pathophysiology, many of them related to lifestyle. In contrast to T2D, morbidity of CHD has been unchanged during the last 10 years, but mortality rates have decreased overall [13].
1.3 Microcirculation
The role of the microcirculation is to adapt perfusion in tissues to local metabolic demands. The microvascular bed is roughly divided into three types of vessels;
mid- and small sized arterioles, capillaries and venules (Fig. 2a), with different functionality corresponding to specific structures and responsibilities. Arterioles are responsible for vascular resistance, venules for collecting blood after passing tissue and capillaries, lacking smooth muscle cells, are responsible for delivery and exchange of molecules in tissue [21]. Heterogeneity in microcirculation allows for adaptations depending on tissue and stimulus, and changing permeability, blood flow and capillary recruitment (the number of blood-perfused capillaries) are all essential parts of a healthy microcirculation [22, 23].
The innermost layer in all vessels consists of a single layer of endothelial cells [24] (Fig. 2b). The endothelium has a gate-keeping role sustaining the integrity of both the circulation and the tissues [24]. Furthermore, the endothelium mediates inflammatory responses [25], and contributes to vascular tone [26], among other functions. The endothelium contributes to the regulation of blood flow and blood pressure by releasing vasodilator and vasoconstrictor substances [27]. The predominant vasodilator released from endothelial cells is nitric oxide (NO) [28,
enzyme NO synthase (NOS) [30]. NO is synthesized by endothelial cells in response to shear stress [31], as well as to insulin [32]. In addition to vasodilation, NO regulates proliferation and inflammation, as well as having oxidative effects [33, 34]. One of the vasoconstrictors produced by the endothelium is the peptide endothelin-1 [35], also induced by insulin [36]. In healthy endothelium, there is a balance between expression of NO and endothelin-1 [37], with low production of endothelin-1 preserving the bioavailability of NO, favouring vasorelaxation [38].
Figure 2. Overview of a microvascular unit (a) and the differences between the anatomies of the vessels (b).
1.4 Insulin resistance and endothelial dysfunction
Insulin resistance demonstrates before overt T2D, and is characterized by decreased sensitivity and/or responsiveness to the metabolic actions of insulin. In an effort to maintain the metabolic effects of insulin, insulin secretion is increased causing hyperinsulinemia [6]. Endothelial dysfunction is proposed to antecede insulin resistance [39], since microvascular dysfunction is a feature seen
endothelial dysfunction is closely related [40], and shares many of the underlying mechanisms, such as glucotoxicity, lipotoxicity and inflammation [41, 42], creating a reciprocal relationship between them. Thus, metabolic and cardiovascular diseases are closely linked together. One of the most consistent observations of vascular function in insulin resistance and T2D is an impaired endothelium-dependent vasodilation across tissues [43], further implicating the prospect of common underlying mechanisms. Moreover, impaired capillary recruitment, decreased blood flow and decreased capillary density (rarefaction) are other ways microvascular dysfunction manifests in insulin resistance (Fig. 3).
Figure 3. Overview of the relationship between insulin resistance and functions of microcirculation with the outcome of decreased skeletal muscle perfusion
The reduced ability of insulin to mediate glucose uptake probably has multifactorial underlying mechanisms. Skeletal muscle blood flow and perfusion have been shown to mediate skeletal muscle glucose uptake [44, 45]. Further, it has been proposed that the importance of the hemodynamic actions of insulin is
which is decreased in insulin resistance [48]. Insulin results in muscle capillary recruitment, which occurs independently of blood flow [49, 50], suggesting a specific action of insulin to induce muscle perfusion. Additionally, structural and functional rarefaction contribute to impaired capillary recruitment in hypertension [51], which might be an explanatory mechanism of the relationship between hypertension and the compromised vascular and metabolic actions of insulin [52]. Further, decreased vascular NO bioavailability has been implicated as a central stimulator of microvessel rarefaction in obese rats [53].
Decreased NO bioavailability is highly involved in endothelial dysfunction [54, 55]. Studies have shown that insulin production of NO is impaired in the insulin resistant state, while in contrast the insulin stimulated production of endothelin-1 remains intact, causing an imbalance between these two molecules [56, 57]. NO and endothelin-1 has been shown to impair each other’s production in cell culture [58, 59], further exhibiting the close relationship between these antagonistic molecules.
1.4.1 Endothelin-1
Endothelin-1 is a 21 amino acid cyclic peptide [60], which acts by binding to endothelin receptors on vascular muscle and endothelial cells [61]. In health endothelin-1 is primarily produced by endothelial cells, however, in pathologies it can be produced by several different cell types [38]. Studies have shown increased circulatory endothelin-1 levels in insulin resistance and T2D [62, 63], in contrast, others did not observe an increase of endothelin-1 in T2D [64].
Preproendothelin is synthesized through transcriptional activation of the preproendothelin gene, which is regulated by nuclear factor-1 (NF-1), activator protein-1 (AP-1) family members, Smad, and GATA-2 [65, 66].
Preproendothelin is cleaved to form big endothelin-1 which then is further cleaved by endothelin converting enzymes (ECE) to form mature endothelin-1 (Fig. 4) [67].
Endothelin-1 is mainly secreted on the basolateral side of endothelial cells [68], where it acts on vascular smooth muscle cells causing vasoconstriction. Two subtypes of receptors mediate the effects of endothelin-1, ETA and ETB [61, 69, 70]. In the vasculature, ETA is mainly located on vascular smooth muscle cells and ETB is primarily located on both vascular smooth muscle and endothelial cells [71], but both receptors are also expressed across other cell types [71]. ETA
and ETB on vascular smooth muscle cells mediate the constrictive effects of endothelin-1, while endothelial ETB on endothelial cells induces vasodilation by
release of prostacyclin and NO [35]. Selective blockade of ETA abolished, while selective blockade of ETB enhanced, the vasoconstrictive effect of endothelin-1 [72], and NO inhibition attenuated the effect of ETA blockade [73]. Further, blockade of ETA improved skin capillary circulation in T2D [74]. Dual receptor antagonism enhanced insulin sensitivity and glucose uptake in obese subjects, and improved endothelial function in T2D and hypertension [75-77]. Finally, chronic endothelin-1 treatment led to insulin resistance in rats [78]. In a Swedish cohort with elderly subjects, circulating endothelin-1 levels did not correlate with endothelium-dependent vasodilation. However, a genotype score constructed from genes involved in the endothelin-1 pathway was significantly correlated to endothelium-dependent vasodilation as well as endothelium-independent vasodilation in forearm resistance vessels in this population [79].
Figure 4. Production of endohelin-1. Released endothelin-1 binds ETA and ETB on smooth muscle cells leading to vasoconstriction. ETB is also present on the endothelial cell where it promotes NO production leading to vasodilation.
1.4.2 Obesity and inflammation
As previously mentioned, obesity is one of the most important risk factors for T2D [80-82]. Obese study subjects show reduced endothelium-dependent vasodilation [83], and reduced transcapillary transport of insulin to skeletal muscle and adipose tissue [84, 85]. A strong association between obesity and disturbances in metabolism has been described, with differences between men and women [82, 86]. However, it should also be noted that insulin resistance is not displayed in all obese [87], and that T2D and cardiovascular risks may be different in metabolically healthy obese.
A chronic low-grade inflammation is described in obesity as well as in T2D [43, 88-90]. Higher circulating levels of inflammatory markers, such as tumour necrosis factor-α (TNF-α), C-reactive protein (CRP) and cytokines have been suggested to contribute to the pathogenesis of T2D [91], and are manifested in overt T2D [92]. Further, TNF-α has been described to impair capillary recruitment and glucose uptake [93, 94]. Adipose tissue in obese mice and humans overexpress TNF-α, which could contribute to insulin resistance [95, 96].
Cross-talk occurs between perivascular adipose tissue and blood vessels [97, 98]
and TNF-α released from perivascular adipose tissue, has been shown to contribute to the imbalance between NO and endothelin-1 [99].
1.5 Insulin signalling and action
The response of endothelial cells to insulin is rapid, and is initiated by insulin binding to the specific insulin receptor leading to phosphorylation of insulin receptor substrates (IRS) [100] and SH2-domain containing protein (Shc), which functions as docking proteins for downstream signalling molecules [42]. Parallel signalling pathways leads to the upregulation of endothelin-1 and NO by insulin [37] (Fig. 5).
The phosphatidylinositol 3-kinase (PI3K) – Akt – endothelial NOS (eNOS) pathway leads to production of NO [101, 102], while the Ras-mitogen-activated protein kinase (MAPK) pathway leads to production of endothelin-1 [103, 104].
The pathways are complex, and includes several feedback loops and crosstalk between the signalling branches [42]. TNF-α induces vasoconstrictor-effects and decrease vasodilator-effects through c-Jun N-terminal kinase (JNK) [105].
Additionally, TNF-α mediated IκB kinase (IKK) [106], among others, can disrupt insulin signalling through serine/threonine phosphorylation of IRS-1.
Figure 5. Insulin signalling in endothelial cells, effects and crosstalk between tissues.
Insulin binds to the insulin receptor (IR) which activates two parallel signalling pathways leading to NO or endothelin-1. NO leads to vasodilation in smooth muscle cells while endothelin-1 leads to vasoconstriction. Insulin signalling in target cells, e.g. adipocytes and skeletal muscle cells leads to translocation of GLUT4 enabling glucose uptake. TNFα from adipose tissue lead to activation of IKK and JNK, which serine/threonine phosphorylates IRS- 1 thereby impeding the PI3K pathway. Free fatty acids (FFA) inhibit the PI3K-pathway in both endothelial cells and skeletal muscle cells. ROS interacts with NO and forms ONOO which leads to an imbalance between NO and endothelin-1. Endothelin-1 impairs IRS-1 in adipocytes and decrease glucose uptake.
Insulin signalling in skeletal muscle and adipose tissue leads to glucose uptake by translocation of glucose transporter 4 (GLUT4) through PI3K [107] (Fig. 5). The cytoplasmic tyrosine kinase portion of the insulin receptor directly interacts with caveolin-1, and both caveolae and caveolin-1 augments the insulin signal and modulates GLUT4 function, thus playing an important role in energy metabolism [108]. Further, impaired insulin signalling in endothelial cells decreased insulin- mediated glucose uptake in skeletal muscle [109]. During oxidative stress, reactive oxygen species (ROS) form in excess (e.g. superoxide or hydrogen peroxide). Superoxide can interact with NO forming peroxynitrite, a highly reactive radical compound, which impedes the insulin signalling by nitrosylation of tyrosine residues on key proteins such as IRS-1 [110]. Further, endothelin-1 is capable of impairing IRS-1 and Shc in cultured adipocytes [111], as well as forming superoxide in arteries [112]. Endothelin-1 also decrease glucose uptake in skeletal muscle in vivo and in vitro, which is reversed by dual endothelin receptor blockade [113, 114].
Thus, dysregulated cross talk between tissues and different cell types are an important factor in the pathology of insulin resistance. This is further illuminated by the notion that free fatty acids inhibits eNOS in endothelium and decrease PI3K activation in skeletal muscle [115, 116], and that depletion of GLUT4 in adipose tissue leads to insulin resistance in skeletal muscle and liver [117].
Moreover, gene and protein expression of IRS-1 in adipocytes is low in subjects at risk of T2D [118] further connecting deficiencies of the insulin signalling to the pathology of T2D.
1.5.1 Transendothelial transport
The endothelial barrier has been postulated to play a role in the development of insulin resistance [119, 120]. Whether transport of insulin is a saturable or non- saturable process is debated [121-123]. However, studies have demonstrated that a functioning insulin signalling is required for insulin uptake and transport by endothelial cells [124, 125]. Further, NO promotes insulin uptake and transport in endothelial cells [126]. Insulin in cultured endothelial cells co-localized with the insulin receptor and caveolin-1 [127, 128], and caveolae has been implicated in the transport of insulin [129]. Further, importance of transendothelial transport has been indicated in vivo [130, 131]. Nonetheless, separating insulin-mediated transport from insulin-mediated capillary recruitment in vivo is challenging [132, 133], and further studies are needed to investigate the involvement of the insulin receptor.
1.5.2 Postprandial vascular response
Patients with insulin resistance and T2D have postprandial hyperglycaemia [134], and an inflammatory response has been reported in the postprandial state [135, 136]. Furthermore, impaired microvascular recruitment and a subsequent reduction of glucose disposal to skeletal muscle are seen after a meal in obesity and T2D [137, 138]. A high-fat meal impaired endothelium-dependent vasodilation [139, 140] in healthy subjects. These postprandial impairments could be explained by an increased oxidative stress following the meal [141, 142], indicating a formation of peroxynitrite. A postprandial increase of triglycerides is seen in obese subjects with metabolic syndrome [143] and potentiates superoxide production by leukocytes [144]. Nocturnal and postprandial free fatty acid release is increased in T2D [145], displaying the unsuppressed lipolysis in insulin resistance.
1.6 Phosphodiesterase-5 inhibition
NO activates soluble guanylate cyclase (sGC) which leads to increased production of cyclic guanosine monophosphate (cGMP) [146] and activation of protein kinase G (PKG) (Fig. 6). Phosphodiesterases (PDEs) are catabolic enzymes grouped into 11 isoenzymes depending on their substrate affinity and selectivity [147]. One of the most studied PDEs is PDE5, which hydrolyses cGMP into its inactive form.
Figure 6. Site of intracellular mechanisms of tadalafil. NO activates sGC, which produces cGMP and activates PKG. PDE-5 hydrolyses cGMP into its inactive form GMP. Tadalafil inhibits PDE-5 leading to a maintained cGMP signalling.
Pharmacological inhibitors of PDE5 have been developed to sustain the NO mediated pathway. Three inhibitors are approved for treatment of erectile dysfunction and pulmonary hypertension - sildenafil, vardenafil and tadalafil.
Sildenafil was the first selective PDE5 inhibitor, and was primary developed for treatment of angina pectoris, although its off-target effects led to the change of
GTP cGMP
PKG GMP
NO
SGC
PDE-5 Tadalafil
focus to erectile dysfunction [148], while tadalafil was developed later [149].
Interestingly, they have emerged as potential pharmaceuticals in treatment of diabetes and cardiovascular diseases. Tadalafil has certain advantages to sildenafil and vardenafil, it has a longer half time of 17.5 h, short time to maximum concentration (2 h) [150], and absorption is not affected by food intake [151, 152]. Additionally, tadalafil is more selective towards PDE5 compared to PDE6 (780:1), as a result visual side effect are rare [153].
Chronic administration of tadalafil has been shown to improve endothelial function and flow-mediated dilation in subjects with erectile dysfunction [154- 157]. Chronic treatment with tadalafil has been reported to decrease circulating levels of endothelin-1 and/or adhesion molecules [156-158] in men with erectile dysfunction, and in systemic sclerosis [159]. Further, tadalafil decreased oxidative stress in cardiomyocytes and circulating inflammatory markers in diabetic mouse models [160, 161]. In contrast, other chronic studies in humans did not observe a decrease of adhesion molecules [162] or an effect on insulin resistance [163]. A study showed increased muscle capillary recruitment and increased glucose conversion to lactate in adipose tissue after acute administration of tadalafil in the fasting state [164]. Similarly, glucose uptake and capillary recruitment in muscle was increased in T2D after acute administration of tadalafil in the fasting state [165], while this was not observed in obese women after an oral glucose load [166]. PDE5 inhibition using sildenafil increased insulin sensitivity in an eNOS knockdown mouse model and improved insulin action in cell culture [167, 168]. PDE5 in endothelial cells is localized in proximity to caveolae, and is involved in modulation of eNOS, suggesting a possible function in insulin action [169].
1.7 The gender aspect
Men and women differs physiologically in terms of body composition and adipose tissue distribution, and abdominal obesity has been shown to be more involved in the development of metabolic disease [170, 171]. Differences between genders have been reported in T2D and cardiovascular diseases, both regarding risk, treatment and outcome. Elderly men have more controlled T2D than elderly women [172]. Women also have more comorbidities and reported lower score on health status and functioning, which could affect their self-care
Endothelial function declines in men after 40 years, while function in women remains stable for 10 more years [174]. Decreased levels of male reproductive hormones in men was associated with higher mortality from cardiovascular disease [175], and increased levels of testosterone in women was associated to a higher prevalence of insulin resistance and CHD [176]. Further, oestrogen appear to have positive effects on endothelial function in both genders, while androgens seem to have more deleterious effects, depending on age and gender of subjects [177]. Hormone replacement therapy has been suggested in some studies to protect against CHD [178], and in other studies contradictory been proposed to increase incidence of CHD [179]. These contrasting results have been suggested to depend on an age-dependent increase of superoxide rather than changed signal mechanisms of oestrogen [180].
1.8 Endothelin-1 as a biomarker and type 2 diabetes heredity
As described above, insulin resistance and obesity can predict the development of future T2D, and endothelial dysfunction is likewise highly implicated in the pathology of T2D and cardiovascular diseases. Increased levels of endothelin-1 was observed in T2D [62], the metabolic syndrome [181] and impaired glucose tolerance (IGT) [182], but whether endothelin-1 could predict T2D remains elusive.
Circulating endothelin-1 is increased in subjects with advanced atherosclerosis and coronary artery disease progression [183]. An observational study showed a predictive role of high circulating endothelin-1 levels during the first indications of atherosclerosis for future cardiovascular events [184]. Additionally, high endothelin-1 levels were associated with CHD in women of all ages, while this observation only was seen in elderly men [185]. This suggests that endothelin-1 might be a predictor of CHD in women.
In a study investigating first-degree relatives of T2D, endothelial dysfunction was associated with insulin resistance [186]. Additionally, microvascular reactivity was impaired in first-degree relatives of T2D [187], and healthy first- degree relatives had insulin resistance and postprandial lipid intolerance [188].
Further, during controlled hyperinsulinemia first-degree relatives had larger subcutaneous fat cells and an increased net lactate release per fat cell [189].
Normoglyceamic offspring of parents with diabetes have defects in insulin
resistance and glucose disposal, more than 10 year prior to the onset of T2D.
Thus, defects in insulin mediated and insulin-independent glucose uptake might precede and predict T2D [190]. Genetic risk scores had a weak but significant association in subjects with parental T2D, and due to the weak association, the authors suggest that shared environmental factors might be a source behind the higher risk of T2D among relatives [191].
2 AIMS
The overall aim of this thesis was to investigate the microcirculation in insulin resistance, with focus on the expression of endothelin-1, through a translational approach.
Paper I To investigate whether circulating endothelin-1 levels predict incident coronary heart disease.
Paper II To investigate whether circulating endothelin-1 levels predict impaired glucose tolerance and type 2 diabetes.
Paper III To investigate whether microcirculation is associated with insulin delivery to subcutaneous adipose tissue and skeletal muscle after an glucose load in type 2 diabetes patients.
Paper IV To investigate whether acute administration of the PDE-5 inhibitor tadalafil could induce positive vascular, metabolic and anti-inflammatory effects after a mixed meal in type 2 diabetes patients.
Paper V To elucidate the molecular action of tadalafil, and investigate the effects of tadalafil in TNF-α stimulated human endothelial cells.
3 PARTICIPANTS AND METHODS
3.1 Paper I and II
3.1.1 Participants
Between 2002-2005 a population-based cohort was established in the municipalities of Vara and Skövde. A random sample (n=2816) stratified by gender and five-year age groups was created based on all residents between 30- 74 years. Criteria for inclusion were: 1. Answering questionnaires. 2. Donating venous blood samples. 3. Completing the physical examination. Participants between 30-50 years were purposely over-sampled by three-fold compared to participants over 50 years. Through 2011-2014 around 2/3 of the baseline population were summoned for a follow-up visit in the same order as at the baseline (n=1834). Of these, 490 declined participation, 35 had moved from the region, 85 had died and 10 were not included for other reasons, leaving 1334 participants in the follow-up.
In all, 2745 participants were included in paper I, and 1099 participants were included in paper II from the Vara-Skövde cohort. A summary of the participants is shown in Table 1.
3.1.2 Study procedure
Participants arrived in the morning after overnight fasting. They signed an informed consent form and venous blood was sampled. Anthropometry was studied with participants in light clothing. In participants without known T2D an oral glucose tolerance test (OGTT) was performed using 75 g glucose to determine their status [192]. During the two-hour wait, participants filled out questionnaires regarding lifestyle. Further, the study nurses collected information about medical history and medication. At the follow-up examination after ten years, the same procedures were repeated.
3.1.2.1 Diagnostic procedures
Standard laboratory tests were performed for serum cholesterol, triglycerides, high-density lipoprotein (HDL)-cholesterol, low-density lipoprotein (LDL)- cholesterol, apolipoproteinB/A1 (apoB/A1), and high sensitivity CRP (hs-CRP).
Estradiol was measured at Unilabs at Skaraborg Hospital in Skövde, and total
testosterone was measured using UniCelTM DxI 800 Beckman Access®
Immunoassay System Main Instrument DxI-1 (Beckman Coulter, Brea, USA).
Blood pressure was measured twice, a minute apart, with the arm adjusted to the heart level, with participants in a resting state. Hypertension was defined in agreement with guidelines [193] and known hypertension was considered as on- going treatment against high blood pressure. Participants without known hypertension were followed up with further blood pressure examinations if the initial diastolic blood pressure was ≥90 mmHg or if systolic blood pressure was
≥140 mmHg. Subjects with three consecutive high readings were diagnosed as being hypertensive.
Coronary heart disease was defined by ICD10 codes as non-fatal myocardial infarction (code I21), percutaneous coronary intervention and/or coronary bypass grafting, or fatal CHD (code I21-23 and I25) [194].
Diagnosis of T2D was confirmed with either two fasting plasma glucose levels
≥7.0 mmol/l or after one 2-hour plasma glucose value ≥11.1 mmol/l after the OGTT and diabetes symptoms [192]. Impaired fasting glucose (IFG) was diagnosed when participants had a fasting plasma glucose level of 6.0-6.9 mmol/l and IGT when 2-hour plasma glucose was 7.8-11.0 mmol/l post OGTT.
Circulating endothelin-1 concentrations was assessed at baseline using Quantiglo Chemiluminescent ELISA for human endothelin-1 from R&D systems (Minneapolis, USA). Serum insulin was analysed using ELISA from DAKO Diagnostics Ltd (Glostrup, Denmark).
Table 1. Characteristics of the participants included in this thesis
Participants Sex (M/F)
Age (years)
BMI (kg/m2)
HOMA-IR
Paper I CHD
Control
52/20 1298/1375
62±11 47±11
27±4 27±5
2.3±1.7 1.6±1.5 Paper II IGT/T2D
Control
56/48 480/515
53±11 46±11
29±4 26±4
1.8±1.0 1.3±0.8
Paper III T2D
Control
3/4 4/2
59±4
59±6 33±2
24±1 2.4±1.1 1.2±1.1 Paper IV T2D tadalafil
T2D placebo
8/2
6/4 58±9
63±5 30±4
31±4 2.3±1.2 2.7±2.3
3.2 Paper III and IV
3.2.1 Participants
In total, 27 participants with T2D and six lean controls were included in two different clinical trials (Table 1). Participants with T2D were excluded if they displayed significant cardiovascular disease, concomitant disease or complications. All included women were postmenopausal. Treatments with β- blockade, glucagon-like peptide-1 (GLP-1) agonists, insulin, dipeptidyl peptidase (DPP)-inhibitors or glitazones led to exclusion in both trials. Additionally, in paper IV, treatment with oestrogens, nitrates, angiotensin converting enzyme (ACE)-inhibitors and glucocorticoids led to exclusion. All participants signed informed consents prior to the start of examinations.
In paper III participants were included if they met the following criteria: Age 50- 70 years for men, and 55-70 years for women; BMI 30-40 kg/m2 for T2D and 18- 25 kg/m2 for lean controls. Participants with T2D were eligible if they were in a good metabolic control with diabetes duration < 5 years. Seven participants with T2D and six age-matched controls were recruited. Participants with treatment for diabetes or hypertension were asked not to take their medication during 10 days prior to their visits, around five days prior to their visits plasma glucose and blood pressure was examined for safety reasons.
In paper IV participants were included if they had a diabetes duration between 3 months and 15 years. Further inclusion criteria were: Age 40-75 years for men, and 50-75 years for women, BMI 25-40 kg/m2 and HbA1c < 57 mmol/mol.
Twenty participants were included in this study. Participants were asked not to take their treatment against diabetes or hypertension for 3 or 7 days, respectively, prior to the study.
3.2.2 Study procedure
In paper III, intramuscular and subcutaneous microdialysis was performed after an overnight fast to investigate interstitial insulin concentrations before and during three hours after an oral glucose load (75 g). Participants were asked to drink the glucose solution during 1-2 minutes. Arterialized venous blood was repeatedly sampled and 133Xenon clearance was used to assess subcutaneous blood flow (Fig. 7). On a second visit, a subcutaneous adipose tissue needle biopsy was performed and endothelial function was assessed with peripheral
Figure 7. Overview of study protocol for paper III.
In Paper IV, intramuscular microdialysis was performed after an overnight fast to investigate interstitial glucose concentrations after acute administration of tadalafil (20 mg) or placebo in combination with a mixed meal (Fig. 8).
Computer-generated randomization of the participants was done 1:1. The randomization was double-blinded, and neither participants nor investigators were aware of which treatment the participants received. The mixed meal was ingested 30 minutes after the participants received their tablet. Total energy content was 786 kcal (3291 kJ); 46% fat, 48% carbohydrates and 7% protein.
Figure 8. Overview of the study design for the randomised controlled trial (RCT) in paper IV.
3.2.2.1 Analytical procedures
Circulating insulin was analysed using Mercodia insulin ELISA. Insulin in dialysates was determined using Mercodia ultrasensitive insulin ELISA (Mercodia, Uppsala, Sweden). Plasma glucose was determined with Hemoque®
Glucose 201+ (Hemoque, Ängelholm, Sweden) (paper III).
In paper IV glucose and urea in dialysates and plasma was determined using a colorimetric method (glucose) and a UV method (urea) on a CMA600 Microdialysis analyser (CMA Microdialysis, Stockholm, Sweden). Arterial Endothelin-1 levels in circulation was determined with Quantiglo Chemiluminescent ELISA for human endothelin-1 from R&D systems. Arterial intercellular adhesion molecule-1 (ICAM-1), vascular cellular adhesion molecule-1 (VCAM-1) and E-selectin was analysed using Human Vascular Injury II kit and Human E-selectin kit from Meso Scale Diagnostics LLC (Rockville, USA). Circulating concentrations of free fatty acids and triglycerides was analysed using spectrophotometric methods at the accredited clinical laboratory at Sahlgrenska University Hospital.
3.2.3 Microdialysis technique
The microdialysis technique is applied for determination of concentrations of interstitial molecules in vivo. In paper III microdialysis was used to measure interstitial levels of insulin in subcutaneous adipose tissue and in skeletal muscle.
In paper IV microdialysis was applied to assess interstitial levels of glucose in skeletal muscle. This is made possible by using thin catheters with a semipermeable membrane. These are inserted into the tissue of interest, and continuously perfused using a microinjection pump allowing for molecules to diffuse and the dialysate is collected and analysed. The relative recovery of a molecule, defined as the exchange between the perfusion fluid and surrounding interstitial fluid, is dependent on several factors; the length of the catheter, molecular cut-off of the membrane, flow rate, temperature, physical and chemical properties of the molecule and the composition of the perfusion fluid [195]. The recovery factor is the ratio between the concentration in the dialysate and the concentration in the interstitial fluid outside the catheter. Interstitial concentrations can therefore be calculated by estimations of the dialysate concentration and the recovery factor.
3.2.3.1 Insulin
Interstitial insulin was measured in periumbilical subcutaneous adipose tissue using two Asahi-catheters with a 3 MDa molecular cut-off [196], and in brachioradialis muscle using two custom-made CMA catheters with dialysis membrane 12 mm x 0.5 mm and a 0.1 MDa molecular cut-off. The catheters were perfused with isotonic saline with 1% albumin and 1.5 mmol/l glucose at a rate of 2.5 µl/min in subcutaneous tissue and 1.5 µl/min in skeletal muscle. The
technique builds on the theory that even though recovery of molecules is different in vitro compared to in vivo the relationship between the recovery of two different molecules could be the same both in vitro and in vivo. Thus, inulin, a polysaccharide, similar to insulin in molecular weight, was injected intravenously as a bolus dose in the beginning of the study day and then continuously infused during four hours, and measured in dialysate during the following hour. Inulin equilibrates completely between plasma and interstitial fluid [198], and can therefore be used to calculate recovery.
Interstitial concentrations of insulin (I-insulin) in subcutaneous adipose tissue and skeletal muscle could then be calculated as:
3.2.3.2 Glucose
Interstitial glucose in the brachioradialis muscle was measured using two CMA catheters 16 mm x 0.5 mm, 20 kDa molecular cut-off, perfused with isotonic saline supplemented with 1.5 mmol/l glucose and 0.5 mmol/l urea at a rate of 2.5 µl/min. Urea was used as a endogenous reference to calibrate the catheters [199].
The relative recovery of urea was estimated from concentrations of urea in dialysate and plasma, and the recovery of glucose could then be calculated by the relationship to urea in vitro [199].
3.2.4 Blood flow measurements
In paper III subcutaneous adipose tissue blood flow (ATBF) was measured using
133Xenon (133Xe) clearance technique [200]. 4-6 MBq of 133Xe (Mallincrodt, Petten, The Netherlands) in gaseous form was injected during 1 minute in the periumbilical subcutaneous adipose tissue at two sites. After 60 minutes of equilibration, the activity at the injection sites was registered in 30-second intervals using GMS411 Mediscint detectors (John Caunt Scientific, Lancashire, England). Subcutaneous blood flow was calculated by using the following formula:
ATBF = − λ ×ln(CA) − ln(CB) nA− nB
×60s ×100g TS
